Statistical Mediation Analysis for Models with a Binary Mediator and a Binary Outcome: the Differences Between Causal and Traditional Mediation Analysis

Abstract

Mediation analysis is an important statistical method in prevention research, as it can be used to determine effective intervention components. Traditional mediation analysis defines direct and indirect effects in terms of linear regression coefficients. It is unclear how these traditional effects are estimated in settings with binary variables. An important recent methodological advancement in the mediation analysis literature is the development of the causal mediation analysis framework. Causal mediation analysis defines causal effects as the difference between two potential outcomes. These definitions can be applied to any mediation model to estimate natural direct and indirect effects, including models with binary variables and an exposure–mediator interaction. This paper aims to clarify the similarities and differences between the causal and traditional effect estimates for mediation models with a binary mediator and a binary outcome. Causal and traditional mediation analyses were applied to an empirical example to demonstrate these similarities and differences. Causal and traditional mediation analysis provided similar controlled direct effect estimates, but different estimates of the natural direct effects, natural indirect effects, and total effect. Traditional mediation analysis methods do not generalize well to mediation models with binary variables, while the natural effect definitions can be applied to any mediation model. Causal mediation analysis is therefore the preferred method for the analysis of mediation models with binary variables.

Paternal Occupation and Delirium Risk in Older Adults: A Potential Marker of Early-Life Exposures

“…Histogram overlaying and smoothing techniques were used to depict continuous (smoothed) distributions of predicted values according to paternal occupation group ( Irizarry & Love, n.d. ). To estimate potential indirect effects due to adulthood SES factors, we tested the mediating effects of participant ADI and income level on the primary associations using the product-of-coefficients method ( Rijnhart et al, 2021). We used point-biserial correlation to measure the collinearity of the two adulthood SES factors.…”

Gender differences in work attendance among health care workers in Northern Nigeria during the COVID-19 pandemic

“…34,35 We first estimated the Natural Direct Effect (NDE) and Natural Indirect Effect (NIE) for binary mediator and outcome and computed the 95% Confidence Interval around NIE using bootstrapping with 5000 replications. 35,36 Finally, we estimated the proportion of total effect between gender and attendance mediated by the burden of household work. 37…”

BMI mediates the association of family medical history with self-reported hypertension and diabetes among older adults: Evidence from baseline wave of the longitudinal aging study in India

“…For example, SRH is a good proxy used for measuring overall health status, however, the self-reporting may be influenced by the subjectivity of respondents. Further, the KHB mediation analysis was not controlled for the effects of exposure-mediator interaction and other interactions on outcome variables ( Coman et al, 2017 ; Rijnhart et al, 2021 ). The causal effects in the mediation modeling along with the other possible pathways in the observed associations need to be considered in future studies.…”